JUNOS Enterprise Switching

Spanning Tree Protocol

The loop issue was detected early on and many switch vendors developed their own loop detection protocols. In 1985, the IEEE set out to develop switching standards that could be used on any vendor, and part of the standard (IEEE 802.1D) included STP.

STP Basics

So, how does STP prevent loops? In simple terms it creates a tree topology with only one available path between the root of a tree and a leaf (Figure 6-2). As the physical switched network evolves, STP reconverges and rebuilds the tree. It does this by electing a single “root” bridge and building paths with this bridge as the starting/ending point.

Note

Although a single root bridge will be elected, the bridge can change over time due to topology or bridge alterations.

The root bridge itself does not prevent loops in our two-switch topology (as in Figure 6-1). The issue is that each bridge forwards a duplicate frame onto the same LAN segment. An easy fix is to simply elect a single switch to forward the frame onto the segment. In STP terminology, a single switch should be elected to forward from the root bridge to any single branch. This ensures that there is only a single path from one end station to another end station. The switch that is responsible for this is called the designated bridge (Figure 6-3).

Note

In this chapter, we use the words switch and bridge interchangeably. Historically, a bridge has two ports and is used to interconnect LANs, usually in software-based devices. A switch is really a multiport ...